Konecta Cuts 11,000 Jobs Betting on AI: The Lesson the Headline Misses
Konecta, Spain's largest contact center group, has executed 11,000 layoffs — most of them in Latin America — while its €945 million syndicated loan trades at a discount of more than 40% on the secondary market. At the same time, the company launched Katalyst 2028, a three-year plan to reach €2.5 billion in revenue powered by generative AI deployed with Google Cloud and Uniphore. The easy headline is "AI replaces people." The headline that should actually worry every contact center leader is different: who audits what the agent replacing those 11,000 people actually does?
What actually happened
| Fact | Figure | Source |
|---|---|---|
| Layoffs executed | 11,000 employees, mostly in Latin America | Merca2, July 2026 |
| Syndicated debt | €945 million, trading at a discount above 40% | Merca2, July 2026 |
| Strategic plan | Katalyst 2028 — target of €2.5B revenue by 2028 | Konecta, official announcement |
| Technology partners | Google Cloud and Uniphore for AI-powered services | Konecta, official announcement |
| Agentic platform | Kolibri, agentic AI moved into production | CMSWire, July 2026 |
The market has read the 44% debt discount as a vote of no confidence in Konecta's ability to refinance in 2029 if the transformation doesn't land in time. That's a reasonable financial read. But from an operational standpoint, there's a more urgent question than solvency: does a company replacing people with AI agents at this speed have the same speed of governance over what those agents actually do in every conversation?
The lesson the headline misses
Replacing people with AI agents is not, by itself, good or bad. It's a workforce decision like any other restructuring. The risk isn't in automating. It's in automating without the same discipline of control that existed — or should have existed — over the human team. A human agent who makes a mistake leaves a trace: a recorded call, a complaint, a supervisor's review. An AI agent that makes the same mistake across thousands of simultaneous conversations can take weeks to show up in any report, if it shows up at all.
This isn't an argument against automating contact centers. It's the argument for auditing that automation with the same rigor that was — or should have been — applied to the human team it replaces. An AI agent in production that misidentifies itself, fails to escalate a real emergency to a human, or hallucinates an answer a customer believes doesn't appear on the operation's KPI dashboard until the damage is already done.
Frequently asked questions
Has Konecta replaced people with AI?
Konecta has executed 11,000 layoffs, mostly in Latin America, alongside the launch of its Katalyst 2028 plan, which bets on generative AI and conversational agents as its growth lever through 2028.
What is the Katalyst 2028 plan?
It's Konecta's three-year strategic plan to reach €2.5 billion in revenue by 2028, built on AI-powered services developed with Google Cloud and Uniphore, including its agentic AI platform, Kolibri.
What's the risk of automating a contact center without auditing the agent?
The main risk is operational blindness: an AI agent can fail to disclose it's AI, fail to escalate a real emergency to a human, or provide incorrect information across thousands of conversations without anyone detecting it until the reputational or regulatory damage is already visible.
How do you audit a Digital Worker in production?
With an independent audit across 100% of real conversations — not a sample, not a test environment — verifying AI disclosure, escalation quality, response accuracy, and regulatory compliance (EU AI Act Art. 50).
Automating without governance vs. automating with independent governance
| Factor | Automating without an audit | Automating with independent governance |
|---|---|---|
| Failure visibility | Discovered once already public or widespread | Detected across 100% of real conversations, before escalating |
| Regulatory compliance | Assumed, not verified | Verified with evidence (EU AI Act Art. 50) |
| Reputational risk | High — the failure is learned from the customer or the press, not the company | Low — the company detects and corrects before it spreads |
| Confidence in the savings | Headcount savings aren't backed by quality evidence | Savings are backed by a verified Trust Score |
Key insight from LEXIC.AI: The savings from replacing 11,000 people with AI agents show up on the balance sheet in the first quarter. The cost of not auditing what those agents say in every conversation shows up later — and almost always in the press.
"Nobody debates whether to automate the contact center anymore. The question companies still aren't asking with the seriousness it deserves is who audits the agent that inherits the customer conversation. Savings without governance aren't savings — they're a risk that hasn't materialized yet," says Sergio Llorens, CEO of LEXIC.AI.
What to do if your contact center is transitioning to AI agents
Before scaling automation across your customer service operation, you need evidence — not an assumption — that the agents already in production disclose correctly, escalate when they should, and don't create regulatory or reputational exposure. An independent Lexic Compass report delivers that verdict on real conversations, not simulated ones.